rends-agent-sdk 1.0.0

Runtime interception SDK for AI agent governance

rends-agent-sdk

Runtime interception SDK for AI agent governance. Wraps agent tool calls with synchronous policy enforcement via the AgentCompliant/Rends platform.

Install

pip install rends-agent-sdk

# With framework adapters
pip install "rends-agent-sdk[langchain]"
pip install "rends-agent-sdk[crewai]"
pip install "rends-agent-sdk[autogen]"
pip install "rends-agent-sdk[all]"

Quick Start

from rends_sdk import RendsClient, RendsClientConfig, ActionContext

async with RendsClient(RendsClientConfig(
    api_key="ac_live_...",
    org_id="your-org-uuid",
    agent_id="your-agent-uuid",
    mode="enforce",  # "enforce" | "monitor" | "dry-run"
)) as client:
    result = await client.govern(
        ActionContext(
            action_type="database_query",
            action_name="select_users",
            resource_type="customer_records",
            input_summary="SELECT * FROM users WHERE region = us-east-1",
        ),
        tool_fn=lambda params: db.query("SELECT * FROM users"),
    )
    # result.allowed, result.status, result.result, result.check.violations

LangChain Integration

from rends_sdk.adapters import govern_langchain_tool

governed = govern_langchain_tool(client, search_tool, resource_type="web_search")
agent = create_react_agent(llm, tools=[governed])

CrewAI Integration

from rends_sdk.adapters import govern_crewai_tool

search = SerperDevTool()
governed = govern_crewai_tool(client, search, resource_type="web_search")
agent = Agent(tools=[governed], ...)

AutoGen Integration

from rends_sdk.adapters import govern_autogen_tool

def search_web(query: str) -> str:
    return tavily.search(query)

governed = govern_autogen_tool(client, search_web, "search_web", resource_type="web_search")
agent.register_function(function_map={"search_web": governed})

Enforcement Modes

Mode	Behavior
enforce	BLOCK → raises GovernanceBlockError, PASS/WARN → executes
monitor	All decisions logged but never block execution
dry-run	Policy checked, nothing executed, nothing logged

Pre-flight Check

check = await client.check_policy(ActionContext(
    action_type="data_export",
    action_name="export_pii",
    resource_type="user_data",
    input_summary="Export all PII to CSV",
))
# check.allowed, check.status, check.violations

Pipeline Stages

The SDK runs a 5-stage pipeline on every govern() call:

1. beforeAction — validates action shape, computes input hash
2. policyCheck — calls POST /v1/compliance/check-action synchronously
3. decisionReturn — parses { allowed, status, violations, warnings }
4. executionGate — executes tool if allowed, raises if blocked (in enforce mode)
5. telemetryEmit — logs to POST /v1/telemetry/ingest (hash-chained audit trail)

Configuration

Field	Type	Default	Description
api_key	str	required	API key (ac_live_...)
org_id	str	required	Organization UUID
agent_id	str	required	Agent UUID
endpoint	str	https://agentcompliant.ai/api	API base URL
mode	str	enforce	Enforcement mode
timeout_ms	int	10000	Policy check timeout
retries	int	2	Retry count for transient failures
fail_open	bool	False	Allow action if policy check fails

License

MIT